I like the idea of 'atomic' but i would call the new components like
CurveHozirontalFrame or CurvePerpendicularFrame 'molecules', since you
can do the same with a couple of other more basic components (same
applies to some older components). Just some thoughts (don't take them
too seriously), maybe they could be like pre-made clusters that you
can explode and get the more basic components. This cluster-components
would be a lot less optimized, i guess, but it would allow non
programmers to create their own custom components...

I had no idea the interval was programming related, but i liked the
idea of have as minimum inputs as possible. Instead of having two
inputs for Max/Min all the time, just handle it in another component.
It relates to the 'hierarchical vs parallel' interface i was
describing above.

On Nov 8, 4:40 pm, David Rutten <[EMAIL PROTECTED]> wrote:
> It's the eternal discussion between splitting and lumping,
> specialization and generalization.
> I chose to make the components as small as possible (i.e. have a
> minimum amount of inputs/outputs), because screen real estate matters
> a great deal on the canvas. It's less important on the toolbars,
> because they have flyouts.
>
> In a platform like Rhino the decision for splitting or lumping is
> purely an esthetic one: "which is easier to learn or understand". In
> Grasshopper it actually has an effect on features as well, if the
> CurveFrame components were merged into 1 rather than 3, you'd be able
> to alter the behaviour of the component using additional Grasshopper
> logic, in the current layout, this is much harder. On the other hand,
> by making the components more 'atomic', readability of a network is
> improved.
>
> The choice I made for the Gradient input (2 numbers instead of 1
> interval) was influenced by the thought that people are more
> comfortable with numbers than intervals (especially when they are new
> to programming), and that probably intervals are rarer than numbers.
> Since it's possible to convert back and forth between intervals and
> number combos without hassle, I picked the most basic approach. Of
> course I /could/ add an option to the Gradient control which would
> switch input parameter types from 2 numbers to 1 interval. I don't
> mind adding options like those to special components, although I'm
> hesitant to add additional settings to regular Components.
>
> --
> David Rutten
> Robert McNeel & Associates
>
> On Nov 7, 6:23 pm, visose <[EMAIL PROTECTED]> wrote:
>
> > - CurveFrame component
> > - CurveHozirontalFrame component
> > - CurvePerpendicularFrame component
>
> > It reminds me of all the different rhino commands that are almost
> > alike. I remember people criticizing this. Maybe there should be a way
> > of adding more options to a component (like the lofting options) to be
> > able to reduce the number of components.
> > I think it's better to have a more hierarchical work flow that having
> > hundreds of "parallel" commands are your disposal. The tabbed
> > interface and sub-categories are a nice way of organizing (imo, rhino
> > should have something similar, the toolbars are not enough) but maybe
> > there should be one more step that reflects in the canvas too.
>
> > The graphs could have inputs for domains (i just comment on this in
> > another thread), and why is the gradient upper/lower limits two
> > different inputs and not an interval?
> > ....I'm trying really hard to think of something else to complain
> > on ....i'll come back later.
>
> > On Nov 7, 4:02 pm, David Rutten <[EMAIL PROTECTED]> wrote:
>
> > > Greetings mere mortals,
>
> > > a minor update to 0.5 is available:
>
> > > Fixes:
> > > - Setting a Graph back to None did not cause an update, this is fixed.
> > > - Graph mapper objects with no graphs failed to read back in, this is
> > > fixed.
> > > - Transformations that do not maintain similarity now properly deform
> > > non-deformable geometry types (circles, arcs and boxes become Nurbs
> > > curves and BReps respectively), this is fixed.
> > > - Orient transformation had a bug which resulted in erroneous mappings
> > > for some input cases, this is fixed.
>
> > > Changes:
> > > - EvaluateCurve component has an extra output for curve span length.
> > > - Curve divisions components have been moved into a separate category.
>
> > > Additions:
> > > - CurveFrame component
> > > - CurveHozirontalFrame component
> > > - CurvePerpendicularFrame component
>
> > > Download the new version 
> > > from:http://download.rhino3d.com/download_rel.asp?rel=412
>
> > > Enjoy,
> > > David
>
> > > --
> > > David Rutten
> > > Robert McNeel & Associates- Hide quoted text -
>
> > - Show quoted text -

Reply via email to